That's a half-truth that comes from the assumption that smaller wheels will always be lighter, but due to lack of engineering or just plain economics smaller wheels rarely achieve this potential. For example, in theory smaller wheels should be able to use fewer spokes, but in practice they never seem to be built that way because standard spoke placement, thicknesses, angles, alloys, etc have all evolved to be optimal for larger wheels and, thus, need to be relatively overbuilt when shoehorned onto small wheel designs (killing any advantage). That's not to say that there aren't small wheeled bikes being built with low spoke counts, it's just that for a given strength it seems that smaller wheels need to use more spokes than they would if optimally designed components were available.

Since the economics of the situation dictates that no manufacturer will custom engineer anything sufficiently close to what is available off the shelf, you have to look to radically different designs to see the effect. For example, an inline speed skater can usually beat the pants off a roadie in sprints or going up hills and, although the mechanics of skating are not as well understood as cycling, due to their re-engineering, inline skate wheels probably exhibit all the purported advantages of smaller wheels.

That's one of the things I find so interesting about being an engineer (albeit not a mechanical one): To engineer a superior product one must engineer a trojan horse that can not only trump existing products, but also weave its way past the status quo. In my opinion, small wheel racing designs have thus far failed to achieve this goal (though I would contend that the 8" wheeled Carryme from Pacific Cycles will blow any other commuter bike out of the water in terms of getting up to speed or going up hills with the minimal amount of effort).

I'm pretty sure they're trying to refer to the rotational inertia, but you can file this under the "don't you have to pedal more?" myth about small wheels because once you account for the higher gearing (and subsequent wheel spinning) the inertia is the same.

I agree, unless he's referring to the (very difficult to measure) energy expended steering the bike it's hard to see how the gyroscopic effect could come into play (which, by the way, Jobst Brandt has argued is negligible compared to the qualitatively similar affect of fork trail).

Well, don't you agree that, neglecting surface friction, it requires twice as much energy to roll a wheel than slide it? Granted, "lift" is a bit misleading as gravity doesn't pull any harder on a rotating mass than a stationary one, but going over a hill is a longer distance than tunneling through it.

Even two different road bikes of different design, but equal weight, are not guaranteed to perform exactly the same. There are too many other variables. The biomechanics of cycling is more complicated than most like to admit.